Author + information
- Received April 4, 2008
- Revision received May 12, 2008
- Accepted May 12, 2008
- Published online September 9, 2008.
- Roberto R. Giraldez, MD,
- Robert P. Giugliano, MD, SM, FACC⁎ (, )
- Satishkumar Mohanavelu, MS,
- Sabina A. Murphy, MPH,
- Carolyn H. McCabe, BS,
- Christopher P. Cannon, MD, FACC and
- Eugene Braunwald, MD, MACC
- ↵⁎Reprint requests and correspondence:
Dr. Robert P. Giugliano, TIMI Study Office, 350 Longwood Avenue, 1st Floor Offices, Boston, Massachusetts 02115
Objectives This study sought to determine whether the benefit of intensive lipid-lowering therapy (LLT) is dependent on baseline low-density lipoprotein cholesterol (LDL-C).
Background Aggressive LDL-C reduction with statins improves cardiovascular outcomes in acute and chronic coronary heart disease (CHD). The importance of baseline LDL-C is unclear.
Methods We compared 2-year composites of death, myocardial infarction (MI), unstable angina, revascularization >30 days, and stroke (primary end point), and CHD death, MI, and revascularization >30 days (secondary end point) in 2,986 statin-naïve patients with recent acute coronary syndrome (ACS) randomized to atorvastatin 80 mg versus pravastatin 40 mg in the PROVE IT–TIMI 22 (Pravastatin or Atorvastatin Evaluation and Infection Therapy–Thrombolysis In Myocardial Infarction 22) study stratified by quartiles of baseline LDL-C. Multivariable models assessed whether the treatment benefit was dependent on baseline LDL-C.
Results A significant reduction in the hazards of the primary (hazard ratio [HR]: 0.63, 95% confidence interval [CI]: 0.47 to 0.85, p = 0.002) and secondary (HR: 0.57, 95% CI: 0.42 to 0.79, p = 0.001) end points occurred in patients within the highest quartile (>132 mg/dl) of baseline LDL-C treated with atorvastatin 80 mg. The benefit of intensive therapy progressively declined as baseline LDL-C decreased. The lowest quartile (LDL-C ≤92 mg/dl) experienced similar rates of the primary (HR: 0.93, 95% CI: 0.69 to 1.25, p = 0.63) and secondary (HR: 0.98, 95% CI: 0.71 to 1.35, p = 0.89) end points. Adjusted interaction tests between treatment and highest versus lowest baseline LDL-C quartile were significant for the primary and secondary end points (p = 0.03 and p = 0.007, respectively). Analyzing baseline LDL-C as a continuous variable, atorvastatin 80 mg was associated with improved outcomes provided the baseline LDL-C was >66 mg/dl.
Conclusions A progressive reduction in the benefit of intensive LLT with atorvastatin 80 mg over pravastatin 40 mg occurred in statin-naïve ACS patients as baseline LDL-C declined. (Pravastatin or Atorvastatin Evaluation and Infection Therapy–Thrombolysis in Myocardial Infarction 22 [PROVE IT–TIMI 22]; NCT00382460)
Lowering of low-density lipoprotein cholesterol (LDL-C) with statin therapy reduces the risk of death and cardiovascular events in both primary and secondary prevention (1). Contemporary trials have extended the findings of the initial studies by showing significant clinical benefit of intensive lipid reduction to much lower LDL-C concentrations (2–4). Currently, the National Cholesterol Education Program (NCEP) recommends an optional target LDL-C of <70 mg/dl in patients at high risk of cardiovascular events (5,6).
Despite the reduction in cardiovascular risk shown with cholesterol-lowering treatment, the PROVE IT–TIMI 22 (Pravastatin or Atorvastatin Evaluation and Infection Therapy–Thrombolysis In Myocardial Infarction 22) trial as well as other analyses in patients with acute and chronic coronary heart disease (CHD) have suggested that this benefit might be limited by baseline cholesterol values, indicating the presence of a threshold below which further LDL-C reduction may not add clinical benefit (2,7,8). Other studies, however, have observed that similar magnitudes of LDL-C reduction translated into fewer cardiovascular events regardless of baseline LDL-C (9–11). Therefore, controversy remains regarding the influence of pre-treatment LDL-C on the clinical benefit of lipid-lowering therapy.
The PROVE IT–TIMI 22 trial showed that intensive statin treatment with atorvastatin 80 mg was associated with a significant reduction in the composite rate of death or major adverse cardiovascular events compared with a standard regimen with pravastatin 40 mg in high-risk patients admitted with acute coronary syndromes (ACS) followed up for an average of 24 months (2). This trial provided the opportunity to evaluate the impact of baseline LDL-C on the benefit of intensive compared with standard lipid-lowering therapy.
Patient population and study protocol
The design of the PROVE IT–TIMI 22 trial has been reported previously (2,12). Briefly, men or women at least 18 years old hospitalized with ACS, either myocardial infarction (MI) (with or without ST-segment elevation) or high-risk unstable angina, in the preceding 10 days were randomly assigned to receive pravastatin 40 mg or atorvastatin 80 mg daily. The protocol required the baseline total cholesterol to be <240 mg/dl in statin-naïve patients or <200 mg/dl in patients on prior statin therapy. Patients were managed with standard medical and interventional treatment for ACS. Major exclusion criteria included treatment with fibric acid derivatives or niacin that could not be discontinued before randomization, or creatinine >2.0 mg/dl.
Patients were followed up for 18 to 36 months (average 24 months) with visits and blood sample collection at 30 days, 4 months, and every 4 months thereafter. Blood samples were analyzed at a central laboratory. Total cholesterol, high-density lipoprotein cholesterol (HDL-C), and triglycerides were measured immediately on freshly shipped plasma samples, using an enzymatic colorimetric method and the Roche Modular system (LabCorp, Raritan, New Jersey), whereas LDL-C was obtained by calculation (Friedewald formula). The dose of pravastatin was increased to 80 mg if the LDL-C level exceeded 125 mg/dl on 2 consecutive measurements. The dose of either study drug was halved in the event of abnormal liver function test results, elevation of creatine kinase levels, or myalgias.
In the current analysis, patients who were not prescribed a statin before the qualifying admission (statin-naïve) were stratified into quartiles of baseline LDL-C obtained at randomization. These patients were selected as the primary cohort of interest because the LDL-C at study entry in this population would more accurately reflect the patients' true baseline level because it was not confounded by prior chronic statin use. The LDL-C at 4 months was considered to be the on-treatment value because it remained largely unchanged throughout the study. The data were analyzed in quartiles of baseline LDL-C to permit multivariate adjustment between subgroups of patients of sufficient size. Results are presented as adjusted hazard ratios (HRs) between subgroups. Baseline LDL-C also was analyzed as a continuous variable to determine a threshold value for the benefit of aggressive lipid therapy.
The primary and secondary outcomes measures in this analysis were the same as for the main PROVE IT–TIMI 22 trial (2,12). The primary end point was the time from randomization to the first occurrence of death from any cause, MI, documented unstable angina requiring hospitalization, revascularization with either percutaneous coronary intervention (PCI) or coronary artery bypass graft (CABG) performed after 30 days following randomization, and stroke. The principal secondary outcome was a composite of death from CHD, nonfatal myocardial infarction, or revascularization after 30 days.
The comparison of baseline characteristics and in-hospital treatment among statin-naïve patients in the different quartiles of baseline LDL-C was performed using the Kruskal-Wallis test for continuous variables and chi-square tests for categorical variables. A Cox regression model was used to evaluate the primary and secondary end points through follow-up (mean 2 years) in the pravastatin 40 mg and atorvastatin 80 mg groups stratified by quartiles of baseline LDL-C. We used Kaplan-Meier estimates to compare the primary and secondary outcomes between treatment groups across LDL-C quartiles. Unadjusted and adjusted models were developed for assessment of an interaction between randomized therapy and quartiles of baseline LDL-C. Multivariable analyses were performed for the prediction of the primary and secondary composite end points using the Cox regression model and included variables of the baseline characteristics significantly associated (p < 0.05) with the outcome of interest in the univariate analysis, treatment assignment, baseline C-reactive protein, and quartiles of baseline LDL-C. Model results are presented as HR with 95% confidence intervals (CIs). A p value <0.05 was the threshold for nominal significance for all comparisons.
The baseline LDL-C was also analyzed as a continuous variable using a locally weighted scatter plot smoother to plot the primary end point using the Logit method. Analyses were performed with Stata/SE version 9.2 (Stata Statistical Software, StataCorp, College Station, Texas).
The analysis cohort included 2,986 statin-naïve patients (72% of the total 4,162 patients) enrolled in the PROVE IT–TIMI 22 trial. Of these, 1,488 patients were randomized to pravastatin 40 mg and 1,498 to atorvastatin 80 mg.
Baseline characteristics stratified by quartiles of baseline LDL-C are shown in Table 1. Patients with lower LDL-C at presentation were older and had a higher prevalence of hypertension and diabetes. In contrast, higher baseline LDL-C values were associated with a history of hyperlipidemia. Despite differences in population characteristics and risks across LDL-C quartiles, Killip class at presentation and rates of in-hospital PCI were similar.
Primary and secondary end points
The composite primary end point at 2 years occurred in 25.4% of the patients in the pravastatin 40 mg group and 20.8% of the patients randomized to atorvastatin 80 mg (HR: 0.80, 95% CI: 0.69 to 0.93, p = 0.004). Similarly, the secondary end point occurred in 22.4% and 18.4% of the pravastatin 40 mg and atorvastatin 80 mg groups, respectively (HR: 0.80, 95% CI: 0.68 to 0.94, p = 0.007).
In the multivariable analysis, the use of atorvastatin 80 mg was an important predictor of both the primary and secondary composite outcomes (Table 2).
Quartiles of baseline LDL-C
The LDL-C quartiles at baseline (on average 7 days after admission) and after 4 months of lipid-lowering treatment with pravastatin 40 mg and atorvastatin 80 mg are shown in Figure 1. Median baseline LDL-C ranged from 81 mg/dl in the lowest quartile to 148 mg/dl in the highest. Statin therapy reduced baseline LDL-C in all quartiles except in the lowest, in which patients treated with pravastatin 40 mg had similar median LDL-C values before and after treatment (from 81 to 82 mg/dl). Absolute reduction in LDL-C from baseline to month 4 values increased progressively and significantly from lowest to highest baseline LDL-C quartiles, regardless of the statin used (Fig. 2). For patients treated with atorvastatin 80 mg, a median 26-mg/dl (34%) decrease in LDL-C level was achieved in the lowest quartile, whereas an 80-mg/dl (54%) decrease in LDL-C was achieved in the highest quartile. Despite greater absolute and relative reductions of LDL-C in patients with higher baseline levels, a stepwise increase in the achieved LDL-C at 4 months was seen from the lowest to the highest quartile of baseline LDL-C regardless of treatment (p < 0.001 for both agents) (Fig. 1). A significant increase from 30 to 40 mg/dl (p < 0.001) in the median absolute difference between achieved LDL-C with pravastatin 40 mg and atorvastatin 80 mg at 4 months was present from lowest to highest quartiles of baseline LDL-C (Fig. 2).
Primary and secondary end points stratified by baseline LDL-C
In the highest baseline LDL-C quartile (>132 mg/dl), the primary end point at 2 years occurred in 29.8% of patients allocated to pravastatin 40 mg, compared with 20.7% in the atorvastatin 80 mg group (HR: 0.63, 95% CI: 0.47 to 0.85, p = 0.002) (Fig. 3). A stepwise decline in the benefit of atorvastatin 80 mg over pravastatin 40 mg was seen from highest to lowest baseline LDL-C quartile (Fig. 3, Table 3). In the lowest quartile (baseline LDL-C ≤92 mg/dl), the primary end point occurred in 23.8% of those receiving pravastatin 40 mg and 21.9% of those treated with atorvastatin 80 mg (HR: 0.93, 95% CI: 0.69 to 1.25, p = 0.63). An interaction test between treatment group and lowest versus highest baseline LDL-C quartile adjusted for differences in baseline characteristics was significant for the composite primary end point (p = 0.03) (Table 3).
When baseline LDL-C was analyzed as a continuous variable, the benefit of atorvastatin 80 mg over pravastatin 40 mg was observed at baseline LDL-C values above approximately 66 mg/dl (Fig. 4) in a multivariate model adjusted for predictors of the primary end point (adjusted interaction p = 0.02).
Intensive lipid-lowering therapy with atorvastatin 80 mg as compared with pravastatin 40 mg also reduced the 2-year event rates of the secondary end point of CHD death, MI, and revascularization in the quartile of patients with the highest baseline LDL-C (HR: 0.57, 95% CI: 0.42 to 0.79, p = 0.001) (Fig. 5), but not in patients within the lowest baseline LDL-C quartile (HR: 0.98, 95% CI: 0.71 to 1.35, p = 0.89). An adjusted interaction test between treatment group and lowest versus highest baseline LDL-C quartile was significant for the secondary end point (p = 0.007) (Fig. 5, Table 3). The relationship between baseline LDL-C as a continuous variable and the secondary end point also suggested the emergence of a benefit in favor of atorvastatin 80 mg compared to pravastatin 40 mg for baseline LDL-C >66 mg/dl (multivariate-adjusted interaction p = 0.006).
Our major finding is that the baseline LDL-C is an important predictor of the benefit of intensive compared with moderate lipid-lowering therapy administered for 2 years in high-risk statin-naïve patients admitted with an ACS. Although a significant reduction in the primary end point of all-cause mortality, MI, unstable angina requiring hospitalization, revascularization performed after 30 days of randomization, and stroke was observed in patients within the highest quartile of baseline LDL-C (median LDL-C 148 mg/dl) randomized to atorvastatin 80 mg compared with pravastatin 40 mg, outcomes were similar between the 2 treatments in patients within the lowest quartile of baseline LDL-C (median LDL-C 81 mg/dl). A test of interaction between statin therapy and lowest versus highest quartile of baseline LDL-C was significant both for the primary and secondary end points, indicating that the magnitude of clinical benefit of atorvastatin 80 mg over pravastatin 40 mg observed in patients with elevated LDL-C at baseline was greater than that for patients in the lowest quartile of baseline LDL-C. Our results showed a decline in the superiority of atorvastatin 80 mg over pravastatin 40 mg from the highest to the lowest baseline LDL-C quartile. When LDL-C was analyzed as a continuous variable, the benefit of intensive lipid-lowering therapy was not seen in patients with baseline LDL-C <66 mg/dl.
The reasons for the reduced benefit of intensive statin therapy in patients with lower baseline LDL-C are not fully understood. Because of the apparent log-linear relationship between LDL-C levels and cardiovascular events, a greater benefit of intensive cholesterol reduction would be expected in patients with elevated baseline LDL-C (5). However, it is also possible that the greater reduction in LDL-C observed with atorvastatin 80 mg in patients with higher baseline values might have partially contributed to this difference in benefit. The absolute difference in LDL-C achieved at 4 months with pravastatin 40 mg versus atorvastatin 80 mg significantly increased by one-third from the lowest to the highest quartile. Disparity in baseline characteristics and hospital treatment among quartiles of baseline LDL-C might have accounted for some of the differences seen in the benefit of intensive therapy as well, despite multivariable adjustments. Patients with lower baseline LDL-C were older and had greater comorbidities, such as diabetes and hypertension, that may have played a more important role in the risk of future events than LDL-C. The benefit of lipid therapy in these groups of patients is still under scrutiny (13,14) although it seems to be a safe strategy (15,16).
Previous studies in patients with acute and chronic CHD have evaluated the role of baseline LDL-C on the benefit of lipid-lowering therapy. Although the LIPID (Long-Term Intervention with Pravastatin in Ischaemic Disease) (8) and CARE (Cholesterol And Recurrent Events) (7) trials showed a lower threshold for the benefit of statin therapy with pravastatin 40 mg, investigators from the 4S study (Scandinavian Simvastatin Survival Study) (9) found a similar benefit of simvastatin over placebo irrespective of baseline LDL-C. Likewise, the Heart Protection Study did not detect a lower boundary for the benefit of simvastatin over placebo (11). Key differences exist between our analysis and earlier placebo-controlled studies that randomized patients with chronic coronary disease and with much higher baseline LDL-C levels at a time when potent lipid-lowering agents were not available. Because of the relatively low mean LDL-C of 112 mg/dl in the statin-naïve population at baseline and highly reduced LDL-C levels achieved with atorvastatin 80 mg in the PROVE IT–TIMI 22 trial, these data provided a unique opportunity to evaluate the impact of baseline LDL-C on the effect of 2 statin regimens of differing LDL-lowering intensities on clinical outcomes in the modern era.
Recent evidence has shown that aggressive reduction of LDL-C to lower levels might be beneficial for the prevention of cardiovascular events (2,3,16–18). However, the optimal target LDL-C remains unknown, and large randomized trials are underway to investigate the benefit of reaching even lower values of LDL-C (19). The markedly reduced LDL-C levels achieved with atorvastatin 80 mg compared with pravastatin 40 mg in each quartile of baseline LDL-C allowed for the comparison of the impact of progressively lower LDL-C on cardiovascular outcomes. In the lowest quartile of baseline LDL-C, cardiovascular events were similar in patients treated with atorvastatin 80 mg and pravastatin 40 mg, although the LDL-C levels achieved after 4 months of therapy were 53 mg/dl and 82 mg/dl, respectively. As achieved LDL-C increased across increasing quartiles of baseline LDL-C, the benefit of more intensive therapy became evident. In the highest quartile, patients with a median LDL-C of 72 mg/dl after 4 months of therapy with atorvastatin 80 mg had improved outcomes compared with those who achieved a median LDL-C of 112 mg/dl after 4 months of pravastatin 40 mg. Therefore, it is possible that reduction of LDL-C to very low levels (<60 mg/dl) in patients with reduced baseline levels before therapy might not be effective in post-ACS patients followed up for 2 years. Our data suggest that most of the overall reduction in cardiovascular events seen for 2 years with intensive lipid lowering in PROVE IT–TIMI 22 was derived from patients with higher baseline LDL-C.
The finding that aggressive lipid treatment might not be required in patients with reduced baseline LDL-C has important clinical implications. The use of moderately intensive regimens in a cohort of patients with reduced baseline LDL-C may result in fewer side effects, reduced drug costs, and higher adherence to therapy (20,21).
The PROVE IT–TIMI 22 trial was not designed with sufficient power to detect differences in treatment effect within quartiles of baseline LDL-C, as was performed in our post hoc analysis. However, a consistent stepwise reduction in the benefit of intensive lipid-lowering therapy was identified from highest to lowest quartile of baseline LDL-C. Additionally, interaction tests showed a difference in the benefit of atorvastatin 80 mg over pravastatin 40 mg in patients with baseline LDL-C >132 mg/dl versus ≤92 mg/dl. One important confounding factor in this analysis is the different baseline characteristics among patients stratified by baseline LDL-C. However, multivariate models that included adjustment for differences in baseline characteristics, baseline C-reactive protein, and an interaction term between therapy and quartiles of baseline LDL-C confirmed a difference in the relative benefit of the 2 regimens of varying LDL-C–lowering intensity. Finally, because the effect of statin therapy was less intense in patients with low baseline LDL-C, it is possible that if even more potent lipid-lowering agents or longer follow-up were used, clinical benefit for the patients in this subgroup might have been observed. This exploratory analysis should be interpreted as hypothesis generating, and needs to be followed by definitive, prospective studies that could ultimately refine recommendations regarding patients with ACS who are most likely to benefit from intensive lipid-lowering therapy. Similarly, the lower threshold of baseline LDL-C of 66 mg/dl observed for the benefit of intensive statin therapy should not be extrapolated for use in clinical practice without further prospective validation.
In this analysis, we showed that in patients after an ACS, baseline LDL-C is an important predictor of the benefit of intensive lipid-lowering therapy compared with therapy of moderate intensity. Over an average follow-up period of 2 years, statin-naïve patients with baseline LDL-C <66 mg/dl did not seem to benefit from a reduction in LDL-C with atorvastatin 80 mg compared with pravastatin 40 mg. This observation suggests that intensive statin therapy may not be necessary in patients whose baseline LDL-C is already low. Indirectly, our data suggest that there may be an actual target LDL-C to reach by lipid-lowering therapies of different intensities. However, further studies will be necessary to address this question. In the interim, our analyses support the current guideline recommendations to consider initiation of statin therapy in high-risk patients with ACS if the LDL-C is above 100 mg/dl, and to target an LDL-C ideally <70 mg/dl (22).
The PROVE IT–TIMI 22 trial was supported by a research grant from Bristol-Myers Squibb.
- Abbreviations and Acronyms
- acute coronary syndromes
- coronary artery bypass graft
- coronary heart disease
- confidence interval
- high-density lipoprotein cholesterol
- hazard ratio
- low-density lipoprotein cholesterol
- myocardial infarction
- National Cholesterol Education Program
- percutaneous coronary intervention
- Received April 4, 2008.
- Revision received May 12, 2008.
- Accepted May 12, 2008.
- American College of Cardiology Foundation
- Cannon C.P.,
- Steinberg B.A.,
- Murphy S.A.,
- Mega J.L.,
- Braunwald E.
- Grundy S.M.,
- Cleeman J.I.,
- Merz C.N.,
- et al.
- Anderson J.L.,
- Adams C.D.,
- Antman E.M.,
- et al.
- Pfeffer M.A.,
- Sacks F.M.,
- Moye L.A.,
- et al.
- Influence of pravastatin and plasma lipids on clinical events in the West of Scotland Coronary Prevention Study (WOSCOPS)
- Wiviott S.D.,
- Cannon C.P.,
- Morrow D.A.,
- Ray K.K.,
- Pfeffer M.A.,
- Braunwald E.
- LaRosa J.C.,
- Grundy S.M.,
- Kastelein J.J.,
- Kostis J.B.,
- Greten H.
- Cannon C.P.,
- Giugliano R.P.,
- Blazing M.A.,
- et al.
- Davidson M.H.,
- Robinson J.G.
- Smith S.C. Jr..,
- Allen J.,
- Blair S.N.,
- et al.